shell sed 用法大全

FILE SPACING:

 

# double space a file

 

sed G

 

# double space a file which already has blank lines in it. Output file

 

# should contain no more than one blank line between lines of text.

 

sed '/^$/d;G'

 

# triple space a file

 

sed 'G;G'

 

# undo double-spacing (assumes even-numbered lines are always blank)

 

sed 'n;d'

 

# insert a blank line above every line which matches "regex"

 

sed '/regex/{x;p;x;}'

 

# insert a blank line below every line which matches "regex"

 

sed '/regex/G'

 

# insert a blank line above and below every line which matches "regex"

 

sed '/regex/{x;p;x;G;}'

 

NUMBERING:

 

# number each line of a file (simple left alignment). Using a tab (see

 

# note on '' at end of file) instead of space will preserve margins.

 

sed = filename | sed 'N;s/

 

//'

 

# number each line of a file (number on left, right-aligned)

 

sed = filename | sed 'N; s/^/ /; s/ *(.{6,})

 

/1 /'

 

# number each line of file, but only print numbers if line is not blank

 

sed '/./=' filename | sed '/./N; s/

 

/ /'

 

# count lines (emulates "wc -l")

 

sed -n '$='

 

TEXT CONVERSION AND SUBSTITUTION:

 

# IN UNIX ENVIRONMENT: convert DOS newlines (CR/LF) to Unix format

 

sed 's/.$//' # assumes that all lines end with CR/LF

 

sed 's/^M$//' # in bash/tcsh, press Ctrl-V then Ctrl-M

 

sed 's/x0D$//' # gsed 3.02.80, but top script is easier

 

# IN UNIX ENVIRONMENT: convert Unix newlines (LF) to DOS format

 

sed "s/$/`echo -e `/" # command line under ksh

 

sed 's/$'"/`echo `/" # command line under bash

 

sed "s/$/`echo `/" # command line under zsh

 

sed 's/$/ /' # gsed 3.02.80

 

# IN DOS ENVIRONMENT: convert Unix newlines (LF) to DOS format

 

sed "s/$//" # method 1

 

sed -n p # method 2

 

# IN DOS ENVIRONMENT: convert DOS newlines (CR/LF) to Unix format

 

# Cannot be done with DOS versions of sed. Use "tr" instead.

 

tr -d <infile >outfile # GNU tr version 1.22 or higher

 

# delete leading whitespace (spaces, tabs) from front of each line

 

# aligns all text flush left

 

sed 's/^[ ]*//' # see note on '' at end of file

 

# delete trailing whitespace (spaces, tabs) from end of each line

 

sed 's/[ ]*$//' # see note on '' at end of file

 

# delete BOTH leading and trailing whitespace from each line

 

sed 's/^[ ]*//;s/[ ]*$//'

 

# insert 5 blank spaces at beginning of each line (make page offset)

 

sed 's/^/ /'

 

# align all text flush right on a 79-column width

 

sed -e :a -e 's/^.{1,78}$/ &/;ta' # set at 78 plus 1 space

 

# center all text in the middle of 79-column width. In method 1,

 

# spaces at the beginning of the line are significant, and trailing

 

# spaces are appended at the end of the line. In method 2, spaces at

 

# the beginning of the line are discarded in centering the line, and

 

# no trailing spaces appear at the end of lines.

 

sed -e :a -e 's/^.{1,77}$/ & /;ta' # method 1

 

sed -e :a -e 's/^.{1,77}$/ &/;ta' -e 's/( *)1/1/' # method 2

 

# substitute (find and replace) "foo" with "bar" on each line

 

sed 's/foo/bar/' # replaces only 1st instance in a line

 

sed 's/foo/bar/4' # replaces only 4th instance in a line

 

sed 's/foo/bar/g' # replaces ALL instances in a line

 

sed 's/(.*)foo(.*foo)/1bar2/' # replace the next-to-last case

 

sed 's/(.*)foo/1bar/' # replace only the last case

 

# substitute "foo" with "bar" ONLY for lines which contain "baz"

 

sed '/baz/s/foo/bar/g'

 

# substitute "foo" with "bar" EXCEPT for lines which contain "baz"

 

sed '/baz/!s/foo/bar/g'

 

# change "scarlet" or "ruby" or "puce" to "red"

 

sed 's/scarlet/red/g;s/ruby/red/g;s/puce/red/g' # most seds

 

gsed 's/scarlet|ruby|puce/red/g' # GNU sed only

 

# reverse order of lines (emulates "tac")

 

# bug/feature in HHsed v1.5 causes blank lines to be deleted

 

sed '1!G;h;$!d' # method 1

 

sed -n '1!G;h;$p' # method 2

 

# reverse each character on the line (emulates "rev")

 

sed '/

 

/!G;s/(.)(.*

 

)/&21/;//D;s/.//'

 

# join pairs of lines side-by-side (like "paste")

 

sed '$!N;s/

 

/ /'

 

# if a line ends with a backslash, append the next line to it

 

sed -e :a -e '/$/N; s/

 

//; ta'

 

# if a line begins with an equal sign, append it to the previous line

 

# and replace the "=" with a single space

 

sed -e :a -e '$!N;s/

 

=/ /;ta' -e 'P;D'

 

# add commas to numeric strings, changing "1234567" to "1,234,567"

 

gsed ':a;s/B[0-9]{3}>/,&/;ta' # GNU sed

 

sed -e :a -e 's/(.*[0-9])([0-9]{3})/1,2/;ta' # other seds

 

# add commas to numbers with decimal points and minus signs (GNU sed)

 

gsed ':a;s/(^|[^0-9.])([0-9]+)([0-9]{3})/12,3/g;ta'

 

# add a blank line every 5 lines (after lines 5, 10, 15, 20, etc.)

 

gsed '0~5G' # GNU sed only

 

sed 'n;n;n;n;G;' # other seds

 

SELECTIVE PRINTING OF CERTAIN LINES:

 

# print first 10 lines of file (emulates behavior of "head")

 

sed 10q

 

# print first line of file (emulates "head -1")

 

sed q

 

# print the last 10 lines of a file (emulates "tail")

 

sed -e :a -e '$q;N;11,$D;ba'

 

# print the last 2 lines of a file (emulates "tail -2")

 

sed '$!N;$!D'

 

# print the last line of a file (emulates "tail -1")

 

sed '$!d' # method 1

 

sed -n '$p' # method 2

 

# print only lines which match regular expression (emulates "grep")

 

sed -n '/regexp/p' # method 1

 

sed '/regexp/!d' # method 2

 

# print only lines which do NOT match regexp (emulates "grep -v")

 

sed -n '/regexp/!p' # method 1, corresponds to above

 

sed '/regexp/d' # method 2, simpler syntax

 

# print the line immediately before a regexp, but not the line

 

# containing the regexp

 

sed -n '/regexp/{g;1!p;};h'

 

# print the line immediately after a regexp, but not the line

 

# containing the regexp

 

sed -n '/regexp/{n;p;}'

 

# print 1 line of context before and after regexp, with line number

 

# indicating where the regexp occurred (similar to "grep -A1 -B1")

 

sed -n -e '/regexp/{=;x;1!p;g;$!N;p;D;}' -e h

 

# grep for AAA and BBB and CCC (in any order)

 

sed '/AAA/!d; /BBB/!d; /CCC/!d'

 

# grep for AAA and BBB and CCC (in that order)

 

sed '/AAA.*BBB.*CCC/!d'

 

# grep for AAA or BBB or CCC (emulates "egrep")

 

sed -e '/AAA/b' -e '/BBB/b' -e '/CCC/b' -e d # most seds

 

gsed '/AAA|BBB|CCC/!d' # GNU sed only

 

# print paragraph if it contains AAA (blank lines separate paragraphs)

 

# HHsed v1.5 must insert a 'G;' after 'x;' in the next 3 scripts below

 

sed -e '/./{H;$!d;}' -e 'x;/AAA/!d;'

 

# print paragraph if it contains AAA and BBB and CCC (in any order)

 

sed -e '/./{H;$!d;}' -e 'x;/AAA/!d;/BBB/!d;/CCC/!d'

 

# print paragraph if it contains AAA or BBB or CCC

 

sed -e '/./{H;$!d;}' -e 'x;/AAA/b' -e '/BBB/b' -e '/CCC/b' -e d

 

gsed '/./{H;$!d;};x;/AAA|BBB|CCC/b;d' # GNU sed only

 

# print only lines of 65 characters or longer

 

sed -n '/^.{65}/p'

 

# print only lines of less than 65 characters

 

sed -n '/^.{65}/!p' # method 1, corresponds to above

 

sed '/^.{65}/d' # method 2, simpler syntax

 

# print section of file from regular expression to end of file

 

sed -n '/regexp/,$p'

 

# print section of file based on line numbers (lines 8-12, inclusive)

 

sed -n '8,12p' # method 1

 

sed '8,12!d' # method 2

 

# print line number 52

 

sed -n '52p' # method 1

 

sed '52!d' # method 2

 

sed '52q;d' # method 3, efficient on large files

 

# beginning at line 3, print every 7th line

 

gsed -n '3~7p' # GNU sed only

 

sed -n '3,${p;n;n;n;n;n;n;}' # other seds

 

# print section of file between two regular expressions (inclusive)

 

sed -n '/Iowa/,/Montana/p' # case sensitive

 

SELECTIVE DELETION OF CERTAIN LINES:

 

# print all of file EXCEPT section between 2 regular expressions

 

sed '/Iowa/,/Montana/d'

 

# delete duplicate, consecutive lines from a file (emulates "uniq").

 

# First line in a set of duplicate lines is kept, rest are deleted.

 

sed '$!N; /^(.*)

 

1$/!P; D'

 

# delete duplicate, nonconsecutive lines from a file. Beware not to

 

# overflow the buffer size of the hold space, or else use GNU sed.

 

sed -n 'G; s/

 

/&&/; /^([ -~]*

 

).*

 

1/d; s/

 

//; h; P'

 

# delete the first 10 lines of a file

 

sed '1,10d'

 

# delete the last line of a file

 

sed '$d'

 

# delete the last 2 lines of a file

 

sed 'N;$!P;$!D;$d'

 

# delete the last 10 lines of a file

 

sed -e :a -e '$d;N;2,10ba' -e 'P;D' # method 1

 

sed -n -e :a -e '1,10!{P;N;D;};N;ba' # method 2

 

# delete every 8th line

 

gsed '0~8d' # GNU sed only

 

sed 'n;n;n;n;n;n;n;d;' # other seds

 

# delete ALL blank lines from a file (same as "grep '.' ")

 

sed '/^$/d' # method 1

 

sed '/./!d' # method 2

 

# delete all CONSECUTIVE blank lines from file except the first; also

 

# deletes all blank lines from top and end of file (emulates "cat -s")

 

sed '/./,/^$/!d' # method 1, allows 0 blanks at top, 1 at EOF

 

sed '/^$/N;/

 

$/D' # method 2, allows 1 blank at top, 0 at EOF

 

# delete all CONSECUTIVE blank lines from file except the first 2:

 

sed '/^$/N;/

 

$/N;//D'

 

# delete all leading blank lines at top of file

 

sed '/./,$!d'

 

# delete all trailing blank lines at end of file

 

sed -e :a -e '/^

 

*$/{$d;N;ba' -e '}' # works on all seds

 

sed -e :a -e '/^

 

*$/N;/

 

$/ba' # ditto, except for gsed 3.02*

 

# delete the last line of each paragraph

 

sed -n '/^$/{p;h;};/./{x;/./p;}'

 

SPECIAL APPLICATIONS:

 

# remove nroff overstrikes (char, backspace) from man pages. The 'echo'

 

# command may need an -e switch if you use Unix System V or bash shell.

 

sed "s/.`echo `//g" # double quotes required for Unix environment

 

sed 's/.^H//g' # in bash/tcsh, press Ctrl-V and then Ctrl-H

 

sed 's/.x08//g' # hex expression for sed v1.5

 

# get Usenet/e-mail message header

 

sed '/^$/q' # deletes everything after first blank line

 

# get Usenet/e-mail message body

 

sed '1,/^$/d' # deletes everything up to first blank line

 

# get Subject header, but remove initial "Subject: " portion

 

sed '/^Subject: */!d; s///;q'

 

# get return address header

 

sed '/^Reply-To:/q; /^From:/h; /./d;g;q'

 

# parse out the address proper. Pulls out the e-mail address by itself

 

# from the 1-line return address header (see preceding script)

 

sed 's/ *(.*)//; s/>.*//; s/.*[:<] *//'

 

# add a leading angle bracket and space to each line (quote a message)

 

sed 's/^/> /'

 

# delete leading angle bracket & space from each line (unquote a message)

 

sed 's/^> //'

 

# remove most HTML tags (accommodates multiple-line tags)

 

sed -e :a -e 's/<[^>]*>//g;/</N;//ba'

 

br>

 

# extract multi-part uuencoded binaries, removing extraneous header

 

# info, so that only the uuencoded portion remains. Files passed to

 

# sed must be passed in the proper order. Version 1 can be entered

 

# from the command line; version 2 can be made into an executable

 

# Unix shell script. (Modified from a script by Rahul Dhesi.)

 

sed '/^end/,/^begin/d' file1 file2 ... fileX | uudecode # vers. 1

 

sed '/^end/,/^begin/d' "$@" | uudecode # vers. 2

 

# zip up each .TXT file individually, deleting the source file and

 

# setting the name of each .ZIP file to the basename of the .TXT file

 

# (under DOS: the "dir /b" switch returns bare filenames in all caps).

 

echo @echo off >zipup.bat

 

dir /b *.txt | sed "s/^(.*).TXT/pkzip -mo 1 1.TXT/" >>zipup.bat

 

TYPICAL USE: Sed takes one or more editing commands and applies all of

 

them, in sequence, to each line of input. After all the commands have

 

been applied to the first input line, that line is output and a second

 

input line is taken for processing, and the cycle repeats. The

 

preceding examples assume that input comes from the standard input

 

device (i.e, the console, normally this will be piped input). One or

 

more filenames can be appended to the command line if the input does

 

not come from stdin. Output is sent to stdout (the screen). Thus:

 

cat filename | sed '10q' # uses piped input

 

sed '10q' filename # same effect, avoids a useless "cat"

 

sed '10q' filename > newfile # redirects output to disk

 

For additional syntax instructions, including the way to apply editing

 

commands from a disk file instead of the command line, consult "sed &

 

awk, 2nd Edition," by Dale Dougherty and Arnold Robbins (O'Reilly,

 

1997; http://www.ora.com), "UNIX Text Processing," by Dale Dougherty

 

and Tim O'Reilly (Hayden Books, 1987) or the tutorials by Mike Arst

 

distributed in U-SEDIT2.ZIP (many sites). To fully exploit the power

 

of sed, one must understand "regular expressions." For this, see

 

"Mastering Regular Expressions" by Jeffrey Friedl (O'Reilly, 1997).

 

The manual ("man") pages on Unix systems may be helpful (try "man

 

sed", "man regexp", or the subsection on regular expressions in "man

 

ed"), but man pages are notoriously difficult. They are not written to

 

teach sed use or regexps to first-time users, but as a reference text

 

for those already acquainted with these tools.

 

QUOTING SYNTAX: The preceding examples use single quotes ('...')

 

instead of double quotes ("...") to enclose editing commands, since

 

sed is typically used on a Unix platform. Single quotes prevent the

 

Unix shell from intrepreting the dollar sign ($) and backquotes

 

(`...`), which are expanded by the shell if they are enclosed in

 

double quotes. Users of the "csh" shell and derivatives will also need

 

to quote the exclamation mark (!) with the backslash (i.e., !) to

 

properly run the examples listed above, even within single quotes.

 

Versions of sed written for DOS invariably require double quotes

 

("...") instead of single quotes to enclose editing commands.

 

USE OF '' IN SED SCRIPTS: For clarity in documentation, we have used

 

the expression '' to indicate a tab character (0x09) in the scripts.

 

However, most versions of sed do not recognize the '' abbreviation,

 

so when typing these scripts from the command line, you should press

 

the TAB key instead. '' is supported as a regular expression

 

metacharacter in awk, perl, and HHsed, sedmod, and GNU sed v3.02.80.

 

VERSIONS OF SED: Versions of sed do differ, and some slight syntax

 

variation is to be expected. In particular, most do not support the

 

use of labels (:name) or branch instructions (b,t) within editing

 

commands, except at the end of those commands. We have used the syntax

 

which will be portable to most users of sed, even though the popular

 

GNU versions of sed allow a more succinct syntax. When the reader sees

 

a fairly long command such as this:

 

sed -e '/AAA/b' -e '/BBB/b' -e '/CCC/b' -e d

 

it is heartening to know that GNU sed will let you reduce it to:

 

sed '/AAA/b;/BBB/b;/CCC/b;d' # or even

 

sed '/AAA|BBB|CCC/b;d'

 

In addition, remember that while many versions of sed accept a command

 

like "/one/ s/RE1/RE2/", some do NOT allow "/one/! s/RE1/RE2/", which

 

contains space before the 's'. Omit the space when typing the command.

 

OPTIMIZING FOR SPEED: If execution speed needs to be increased (due to

 

large input files or slow processors or hard disks), substitution will

 

be executed more quickly if the "find" expression is specified before

 

giving the "s/.../.../" instruction. Thus:

 

sed 's/foo/bar/g' filename # standard replace command

 

sed '/foo/ s/foo/bar/g' filename # executes more quickly

 

sed '/foo/ s//bar/g' filename # shorthand sed syntax

 

On line selection or deletion in which you only need to output lines

 

from the first part of the file, a "quit" command (q) in the script

 

will drastically reduce processing time for large files. Thus:

 

sed -n '45,50p' filename # print line nos. 45-50 of a file

 

sed -n '51q;45,50p' filename # same, but executes much faster

 

If you have any additional scripts to contribute or if you find errors

 

in this document, please send e-mail to the compiler. Indicate the

 

version of sed you used, the operating system it was compiled for, and

 

the nature of the problem. Various scripts in this file were written

 

or contributed by:

 

Al Aab <[e-mail]> # "seders" list moderator

 

Edgar Allen <[e-mail]> # various

 

Yiorgos Adamopoulos <[e-mail]>

 

Dale Dougherty <[e-mail]> # author of "sed & awk"

 

Carlos Duarte <[e-mail]> # author of "do it with sed"

 

Eric Pement <[e-mail]> # author of this document

 

Ken Pizzini <[e-mail]> # author of GNU sed v3.02

 

S.G. Ravenhall <[e-mail]> # great de-html script

 

Greg Ubben <[e-mail]> # many contributions & much help